Climate Change and Tropical Andean GlacierRecession: Evaluating Hydrologic Changes

and Livelihood Vulnerability in the CordilleraBlanca, Peru

Bryan G. Mark,∗ Jeffrey Bury,† Jeffrey M. McKenzie,‡ Adam French,† and Michel Baraer‡

∗Department of Geography and Byrd Polar Research Center, The Ohio State University†Environmental Studies Department, University of California, Santa Cruz

‡Earth and Planetary Sciences, McGill University

Climate change is forcing dramatic glacier mass loss in the Cordillera Blanca, Peru, resulting in hydrologictransformations across the Rio Santa watershed and increasing human vulnerability. This article presents resultsfrom two years of transdisciplinary collaborative research evaluating the complex relationships between coupledenvironmental and social change in the region. First, hydrologic results suggest there has been an averageincrease of 1.6 (± 1.1) percent in the specific discharge of the more glacier-covered catchments (>20 percentglacier area) as a function of changes in stable isotopes of water (δ18O and δ2H) from 2004 to 2006. Second,there is a large (mean 60 percent) component of groundwater in dry season discharge based on results from thehydrochemical basin characterization method. Third, findings from extensive key interviews and seventy-tworandomly sampled household interviews within communities located in two case study watersheds demonstratethat a large majority of households perceive that glacier recession is proceeding very rapidly and that climatechange–related impacts are affecting human vulnerability across multiple shifting vectors including access towater resources, agro-pastoral production, and weather variability. Key Words: climate change, glacier recession,hydrology, livelihoods, vulnerability.

El cambio climatico esta causando una dramatica perdida de masa en los glaciares de la Cordillera Blanca, Peru,generando transformaciones hidrologicas en toda la cuenca del Rıo Santa e incrementando la vulnerabilidadhumana. Este artıculo presenta los resultados de dos anos de investigacion colaborativa transdisciplinaria paraevaluar las complejas relaciones entre los cambios ambientales y sociales en la region. Primero, los resultadoshidrologicos sugieren que ha habido un incremento promedio del 1.6 (±1.1) por ciento en la descarga especıficade los desagues con mayor cobertura de glaciares (>20 por ciento de area glaciada), como una funcion del cambioen isotopos estables del agua (δ18O and δ2H) entre 2004 y 2006. Segundo, hay un gran componente (mediade 60 por ciento) de agua subterranea en el descargue de la estacion seca basado en resultados del metodo decaracterizacion de la cuenca hidroquımica. Tercero, los descubrimientos derivados de detalladas entrevistas ainformantes claves y setenta y dos entrevistas de muestra aleatoria administradas a hogares de comunidadespertenecientes a dos estudios de caso de cuencas, demuestran que la gran mayorıa de la gente intuye que larecesion de los glaciares esta avanzando rapidamente y que los impactos relacionados con cambio climaticoafectan la vulnerabilidad humana por medio de muchos vectores cambiantes, incluyendo el acceso a los recursoshıdricos, produccion agro-pastoral y variacion meteorologica. Palabras clave: cambio climatico, recesion de glaciares,hidrologıa, medios de vida, vulnerabilidad.

Annals of the Association of American Geographers, 100(4) 2010, pp. 794–805 C© 2010 by Association of American GeographersInitial submission, July 2009; revised submission, November 2009; final acceptance, January 2010

Published by Taylor & Francis, LLC.

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

Climate Change and Tropical Andean Glacier Recession 795

Rapid glacier recession in the tropical PeruvianAndes due to recent climate change exempli-fies environmental changes in mountain regions

that have critical water supply implications for mil-lions of people globally (Intergovernmental Panel onClimate Change 2007; Viviroli et al. 2007). Modernglacier recession is strongly correlated with a signifi-cant rising trend in atmospheric temperatures, and cli-mate models predict enhanced future rates of glacierloss with increased warming at higher tropical altitudes(Barry and Seimon 2000; Bradley et al. 2009). This re-cession will significantly affect the future availabilityand use of water resources because human populationsare highly dependent on glacial melt water to bufferthe seasonally arid climate of the central Andes (Vuilleet al. 2008). Although progress has been made in trac-ing glacier changes over time and space, new and moreintegrated research is needed to evaluate these ongoingsocial and hydrological transformations (Mark 2008).

The Cordillera Blanca, Peru, is a unique location tostudy these issues as more than 25 percent of the world’stropical glaciers are located in the range. There is alsoa long history of environment–society dynamics in thisregion that has been densely settled for many centuriesand the spatial extent of human activities is often con-ditioned by water availability. The range has also beenthe epicenter of some of the most disastrous hazards inrecorded history (Carey 2010). Finally, the differentialglacial coverage of watersheds in the Cordillera Blancaregion provides a natural continuum to compare hy-drologic and human responses to climate change acrosssubcatchments simultaneously as a proxy for sequentialchanges in time. This article presents results from ourongoing collaborative research project in the region.Working as a transdisciplinary team, we are evaluat-ing the regional impact of glacier melt on seasonal andinterannual water availability and assessing human vul-nerability and governance shifts related to the politicaleconomy of climate change.

Setting

The Cordillera Blanca contains the world’s largestconcentration of tropical glaciers, most of whichflow westward toward the Santa River (Figure 1 andTable 1). According to glacial inventories conducted inthe 1970s, 35 percent of Peru’s glacierized area and40 percent of its glacial reserves (by volume) are storedin the more than 700 glaciers located in the range

Table 1. Characteristics of the Cordillera Blancawatersheds that have available historical discharge data

Watershed PercentageStream Period of record area (km2) glaciated area

Llanganuco 1953–1997 85 36Chancos 1953–1997 221 22Quilcay 1953–1983 243 17Olleros 1970–1997 175 10La Balsa 1954–1997 4,768 8Pachacoto 1953–1983 194 8Quitaracsa 1953–1997 383 7Querococha 1956–1996 62 3

Note: The case study watersheds are in bold type, and the La Balsa dischargestation is italicized to indicate that it is located on the Rio Santa.

(Ames et al. 1989). The Santa River flows northwestover 300 km, draining a total watershed of 12,200 km2

to become the second largest and most regular flow-ing Peruvian river to reach the Pacific Ocean (Markand Seltzer 2003). The upper Santa River watershed,or Callejon de Huaylas, has an area of 4,900 km2 de-fined by the outflow at the La Balsa station (Figure 1).Within the Callejon de Huaylas, tributary streams tothe Santa River from the Cordillera Blanca define sub-catchments with different percentages of glacial cover-age. For some of these tributaries historical dischargerecords are available for forty years. Monthly averagedischarge from these gauged tributary streams is higherduring the months of October through April, closelyreflecting the seasonality of precipitation where greaterthan 80 percent of precipitation falls between Octo-ber and May, and the austral winter months of June toSeptember are known as the dry season.

According to our subdistrict spatial analysis ofthe 2007 Peruvian Census data, the population ofthe Callejon de Huaylas is approximately 267,000inhabitants. Major urban and periurban clusters alongthe valley include the cities of Huaraz (96,000), Caraz(13,000), Yungay (8,000), Carhuaz (7,200), Recuay(2,700), and Catac (2,400), and approximately 1,500small rural settlements (INEI 2007). Our populationestimate is roughly 25 percent higher than previousestimates, which significantly heightens the potentialsocial risks of recent climatic change in the region(Byers 2000; Young and Lipton 2006). The watershedis also a critical source of water for urban centers,agricultural activities, and several hydroelectric powerplants that account for approximately 10 percent of the

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

796 Mark et al.

Figure 1. Map of the Cordillera Blanca,with study sites indicated. Inset tablesinclude data on glacier coverage (Ameset al. 1989; Racoviteanu, Arnaud, andWilliams 2008).

country’s hydroelectric capacity (Ministry of Energyand Mines 2008).

Livelihoods in the region are generally dependent onaccess to water and other natural resources for agricul-tural and livestock production, as more than 80 percentof the population of the region is engaged in small-holder production (INEI 2007). Since the late 1990s,agricultural production along the Santa River has beenshifting toward water-intensive crops for consumptionoutside the region and, similar to trends across the coun-try, several new transnational polymetallic mining facil-ities have recently begun operating along the CordilleraNegra (Bury 2005; Bebbington and Bury 2009). How-ever, more than 50 percent of the population still livesin conditions of poverty (defined as lacking more than

one basic necessity) and 33 percent of the rural popu-lation is illiterate (INEI 2007).

Background and Rationale

Observations of continued glacier recession existthroughout the Andes, but the Cordillera Blanca is oneof only a few locations in the tropics where researchhas quantified changes in the hydrologic regime due toglacier volume loss on a scale relevant to human im-pact. The range has lost one third (>30%) of its glacier-ized area since maximum extensions of the nineteenthcentury, with clear evidence of significant climatechange over the twentieth century (Vuille et al. 2008).Multiscale studies have traced climate forcing and the

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

Climate Change and Tropical Andean Glacier Recession 797

hydrologic impacts of glacier volume changes over time(Mark 2008). Mark and Seltzer (2003) estimated that35 to 40 percent of the mean annual discharge in theQuerococha watershed during 1998 and 1999 was sup-plied by nonrenewed melt from the Yanamarey glacier.This increased to 58 percent over the observation pe-riod from 2001 to 2004 (Mark, McKenzie, and Gomez2005). The Yanamarey glacier is projected to disappearwithin a decade and our recent research has been di-rected toward understanding the implications of thislate-stage glacial retreat for hydrologic processes anddownstream communities (Bury et al. forthcoming).

Changes in the stable isotopes of water observed overrecent years in glacier-fed streams demonstrate relativeincreases in glacier melt water (Mark and McKenzie2007). However, this enhanced melt contribution willdiminish as glacier mass disappears, causing the streamsto have smaller dry season flow and increased variabil-ity (Kaser et al. 2003; Mark and Seltzer 2003; Juen,Kaser, and Georges 2007). Glacier recession also modi-fies watersheds by forming wetlands, lakes, and ground-water reservoirs that alter the surface drainage. Thereis an outstanding need to understand and quantify hy-drologic processes in these dynamic and transforminglandscapes, particularly in view of the potentially severewater stress impacts of glacier loss highlighted in futureclimate change scenarios for Peru.

Complementing research on glacier recession andhydrologic change, our research also builds on recentenvironment–society research examining the adaptivebehavior of human communities to environmentalchange in the Peruvian Andes (Young and Lipton 2006;Postigo, Young, and Crews 2008). Geographic researchin the Andes has long focused on human responsesand adaptation to diverse processes of environmentaland socioeconomic change (Sauer 1941; Knapp 1991;Zimmerer 1991; Denevan 1992; Bebbington 2000; Bury2004). As Sauer (1941) noted, the complex shifts thatclimatic change induces across landscapes have been anenduring theme in geography.

This research also has theoretical links to researchon human vulnerability and adaptation to globalchange. In geographic studies, vulnerability researchhas evolved significantly from its foundations in earlyrisk-hazard studies (White 1942; Burton, White, andKates 1978) and subsequent critiques of this work fora lack of attention to political economy and structuralconditions (Hewitt 1983; Watts and Bohle 1993). Morerecent vulnerability studies have developed integra-tive approaches that examine how biophysical, social,economic, and political factors interact with and feed

back on one another across scales of analysis (Liverman1990; Cutter 2003; Turner et al. 2003; Polsky 2004;Eakin 2006; Leichenko and O’Brien 2008). Much ofthis work makes a case for the transdisciplinary collab-orations, place-based analyses, and mixed methodolog-ical approaches that we are currently utilizing in ourresearch.

Methodology

The linked foci of this research are on shifts in re-gional water resources and the impacts of such changeson household livelihood strategies and vulnerabilityin the Cordillera Blanca. To examine these questionswe have integrated in situ observations with geospatialanalyses, hydrochemical mixing models, semistructuredhousehold surveys, and key interviews. Our methodol-ogy also identifies and analyzes patterns across nestedspatial and temporal scales. The uneven distribution ofmodern glaciers in the tributary watersheds of the SantaRiver provides a natural continuum over which to eval-uate changes in hydrologic processes related to differentamounts of glacier coverage. Furthermore, the researchis focused on the dry season when water resources aremore stressed and glacier melt production is relativelymore important.

Using geographic information systems, remote sens-ing techniques, and site visits we selected three rep-resentative tributary watersheds to the Santa Riverwith different glacial coverage, variable hydrologicalcharacteristics and diverse livelihood pursuits to un-derstand and measure hydrologic processes, calibratehydrochemical mixing models, and evaluate the spa-tial distribution of household resource use patterns.These case study watersheds are Llanganuco, Quero-cocha, and Quilcay (Figure 1). Specific digital productsutilized to identify glacierized areas, generate digitalelevation models, delimit watershed areas, and eval-uate human activities include satellite imagery fromthe Advanced Spaceborne Thermal Emission and Re-flection Radiometer (ASTER), Landsat, and Quickbirdplatforms.

Historic discharge records in the Callejon de Huay-las are available for the Santa River at La Balsa andfor eleven tributaries of the Santa River, dating fromthe 1950s through the end of the twentieth centurywhen observations were largely disbanded due to theprivatization of the state-run water resources office(Mark and Seltzer 2003). To compare trends overtime, normalized anomalies of annual discharge were

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

798 Mark et al.

generated for La Balsa and seven glacierized, noncon-trolled (i.e., no dams) tributary watersheds with thelongest records (Table 1). Discharges were normalizedby subtracting the series mean from the annual value,then dividing that difference by the series standard devi-ation. We computed anomalies for both the total annualdischarge and the dry half of the year (May–October).

To characterize the temporal and spatial variabilityof glacier melt water discharge in the case study water-sheds, we measured dissolved ion concentrations andisotopic signatures (δ18O and δD) of water samples col-lected between 1998 and 1999 and 2004 and 2009 (afterMark and Seltzer 2003; Mark and McKenzie 2007). Weused the hydrochemical basin characterization method(HBCM), based on a multicomponent mass balance ap-proach to identify hydrologic end member’s relative orabsolute contributions at a sample point (Baraer et al.2009). Dry season discharge measurements and hydro-chemistry results from 1998 and 1999 and 2004 to 2008at the Querococha watershed were used to calibratethe HBCM and examine interannual variations of themelt water contribution from a single glacier (Markand Seltzer 2003; Baraer et al. 2009). Spatial variabilityof glacier melt water was evaluated based on samplescollected in the case study watersheds, spread alongthe Cordillera Blanca, during the 2008 dry season. Asimplified two-component hydrograph separation usinghydrochemical and isotopic signatures was used to de-convolve watershed discharge into melt water andgroundwater. The interannual changes in water iso-topes from tributary streams to the Rio Santa were alsoanalyzed to trace the relative changes in glacier meltcontribution to the larger Callejon de Huaylas water-shed (after Mark and McKenzie 2007).

We formulated a mixed set of social science methodsdirected toward maximizing levels of objective confi-dence and minimizing potential biases in our findings.We purposively selected case study communities in eachwatershed to generate the largest set of observable liveli-hood activities and possible impacts of recent glacierrecession. We developed a stratified random samplingframe for individual household selection based on pre-liminary participatory community mapping activities.Individual case study households were then selected inthe field based on their proximity to randomly gen-erated coordinates using portable Global PositioningSystem field units and 1 m resolution satellite imagery(Figure 2). We conducted intensive semistructuredhousehold surveys and a diverse array of unstructuredkey interviews in each community. Overall, the find-ings reported in this article are based on seventy-two

Table 2. Demographic variables listed for both theQuerococha (QUER) and Quilcay (QUIL) case study

watersheds

Case study sample demographic summary QUER QUIL

Total case study area population 3,249 1,200Total households in sample population 40 32Total sample population 181 124Household sample percentage of total community

population5.5 10.3

Average age of respondents 47 51Gender percentage of respondents

Male 35 53Female 65 47

semistructured household surveys, twenty-one unstruc-tured household surveys, and thirty-seven formal keyinterviews in the Querococha (QUER) and Quilcay(QUIL) case study watersheds (Table 2).

Results

Shifts in Regional Water Resources

Glacier-fed stream discharge from the CordilleraBlanca correlates strongly with climate changes, butthe magnitude of glacier melt influence is scale de-pendent. Streams draining the two watersheds withgreatest amounts of glacier coverage (glacial > 20% ofwatershed by area, n = 2) experienced a significant in-crease (p = 0.023) in average annual discharge over theforty-three-year period of historical records (Figure 3).However, there is no significant trend in annual dis-charge averaged for all tributaries with glaciers (n = 7),implying that glacier melt enhancement is not discern-able on an annual basis below a threshold amount ofglacier coverage. A significant (p = 0.004) correlationbetween annual discharge and regional mean air tem-perature over the same time suggests that streamflowresponds rapidly to regional-scale climatic forcing.

In the same glaciated watersheds, dry season (May–October) stream discharge increased significantly untilthe early 1980s but since 1983 has declined considerably(Figure 4). This change in trend indicates that glaciermelt water buffers discharge only temporarily in localwatersheds, and future dry season streamflow is likelyto decline further. On the Santa River scale, dischargedraining the entire Callejon de Huaylas watershed (<10 percent glacier coverage) at La Balsa has declined

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

Climate Change and Tropical Andean Glacier Recession 799

Figure 2. Shaded relief maps of portions of the populated sections of the Quilcay (1) and Querococha (2) case study watersheds, illustratingthe location and distribution of case study locations and glaciers. Digital elevation data obtained by airborne light distance and ranging(LIDAR) survey flown in 2008 (by Horizons South America, S.A.C.).

Figure 3. Average normalized anomalies of annual discharge forthose tributary watersheds of the Santa River with > 20 percentglacier coverage (n = 2; see Table 1) and annual deviation of tem-perature from the 1961–1991 average from twenty-nine Peruvianmeteorological stations between 9–11◦S (from Mark and Seltzer2005).

very significantly (p = 0.004), equaling a 17 percentdecline from 1954 to 1997 (Figure 5).

Our analysis of stable isotope values from Callejonde Huaylas subcatchments with and without glaciersconfirms a relative increase in dry season dischargedue to recent glacier melt (Figure 6). Annual dryseason samples (2004–2008) from Cordillera Blancatributaries with glaciers describe a progressive deple-tion trend in δ18O (i.e., more negative), whereasnonglacierized Cordillera Negra waters show no system-atic change. The total isotopic change correlates to wa-tershed glacier coverage (area percentage), translatingto an increase in area-averaged discharge of 1.6 (± 1.1)percent (Mark and McKenzie 2007).

Relative groundwater contributions during the dryseason were first evaluated using HBCM at Q2, theinflow to Querococha Lake (Figure 2), to explore in-terannual variability. This location drains a 28 km2

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

800 Mark et al.

−3

−2

−1

0

1

2

3

4

1953 1957 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997

No

rmal

ized

dis

char

ge

ano

mal

y(m

s)

3-1

Figure 4. Normalized discharge anomalies for glacierized tributaries(n = 7) of the Santa River from 1953 to 1997. Prior to 1983, there isa significant (p = 0.025) positive trend shown with dark line; after1983 there is a downward trend that is not quite significant (p =0.057) shown in lighter gray. Vertical bars show range of variabilityat 2σ level (± 1 SD).

subwatershed with 7.2 percent glacier coverage. Re-sults indicate a median 59 percent dry season ground-water contribution to the Q2 discharge (Figure 7).The largest calculated contribution was 74 percent in2007 and the minimum contribution was 18 percent in1998. With such high variability and dominant over-all contribution, groundwater is likely as essential asmelt water in Q2 discharge. Second, HBCM appliedto different tributary watersheds simultaneously indi-cates that groundwater aquifers having significant yieldlikely exist in each case study watershed. A simplifiedtwo-component model of the 2008 dry season estimates

−3

−2

−1

0

1

2

3

4

1953 1957 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997

No

rmal

ized

dis

char

ge

ano

mal

y(m

s)

3−1

Figure 5. Normalized anomaly of annual discharge for La Balsastation on the Santa River (1953–1997).

Figure 6. Oxygen isotope values of annual dry season water sampledfrom three watersheds with varying amounts of glacier coverage(dotted lines with symbols) plotted with average isotopic values forthe Cordillera Blanca (with glaciers) and the Cordillera Negra (noglaciers) from 2004 to 2008, shown in thick lines with 2σ variabilityin vertical lines (± 1 SD).

groundwater contributes 0.19 m3s–1 (77 percent) of dis-charge at Q2 and 0.15 m3s–1 (26 percent) of Quil-cay discharge. The difference between these estimatedgroundwater contributions is partly explained by therelative amount of glacier coverage, equal to 3 and 17percent in the Querococha and Quilcay watersheds,respectively.

Livelihoods and Emerging Vectors of Vulnerability

Communities in the Querococha and Quilcay wa-tersheds are embedded within diverse communal, pri-vate, and public governance institutions that affecthousehold access to key livelihood resources such asland and water. Land tenure and land use vary acrosswatersheds and among a variety of institutions begin-ning with privately titled land parcels at the bottomsof the watersheds, ranging upward across communallymanaged lands and terminating at high lakes and theglaciers that remain the sole property of the state andare within Huascaran National Park. Access to the up-per watersheds is thus governed by a complex and of-ten conflictual nexus of state–community regimes (seeFigure 2). For example, grazing access to the upper Que-rococha watershed is controlled by the 2,200 membersof the Campesino Community of Catac and access tothe upper Quilcay watershed is controlled by the 240associates of the Quilcayhuanca Users Group. In ad-dition, a number of land management practices acrossboth watersheds are officially proscribed by NationalPark authorities.

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

Climate Change and Tropical Andean Glacier Recession 801

0

10

20

30

40

50

60

70

80

1998 1999 2004 2005 2006 2007

Gro

un

dw

ater

co

ntr

ibu

tio

n (

%)

Figure 7. Interannual variation ofthe groundwater relative contributionmodeled by hydrochemical basin char-acterization method at the inflow toQuerococha Lake, Q2, in the Quero-cocha watershed.

Households engage in a variety of livelihood activ-ities across both watersheds, but the principal sourcesof household subsistence are agriculture and livestock.Table 3 presents a comparative summary of householdlivelihoods across both case study communities. House-holds in Catac are more dependent on livestock pro-duction largely due to the fact that they have access

Table 3. Major livelihood activities in the Querococha(QUER) and Quilcay (QUIL) case study watersheds

Percentage ofhouseholds

Type of activity Description and variety QUER QUIL

Livestock Cattle, sheep, pigs, horses,guinea pigs, burros

85 78

Agriculture Potatoes, olluco, oca, mashua,corn, wheat, barley, oats,quinua, beans, herbs

68 100

Agroforestry Eucalyptus, pine, quenual(Polylepis), capuli (Prunusserotina), colle (Buddlejacoriacea)

68 72

Commercial Livestock, agricultural products,market, prepared food

45 78

Manual labor Temporary agricultural labor,carpentry, manufacturing

38 40

Tourism Guiding, animal caretaker, burrorental, cook, boat rental

25 27

Dairy products Milk, cheese, eggs 15 9Artisanal Hand-spun fabrics, clothing,

ceramics15 3

to much more communal land (∼66,000 ha) and thatthe lowest vertical range of the community (3,500 mabove sea level) limits the production of key staplecrops such as corn. Conversely, households in the Quil-cay watershed engage in more agricultural, commercial,and manual labor activities because part of the commu-nity is located at a lower elevation (3,300 m above sealevel) and due to the proximity of Huaraz, where manyhousehold members travel on a frequent basis to gener-ate income.

Our evaluation of the ways in which household liveli-hood vulnerability is being affected by climate changeand glacier recession demonstrates the combined waysin which new vectors of environmental and socialchange have begun to affect household activities andaccess to resources. Table 4 presents a summary of find-ings on household perceptions of these changes. Nearlyall of the households in the case study communities areacutely aware of the pace of glacier recession that is tak-ing place in nearby watersheds and frequently noted ingreat detail the accelerating rate and magnitude of re-cession that has been taking place over the past severaldecades. In addition, older respondents often presenteddetailed accounts of the extent of glacial coverage in theupper watersheds long before instrumental records werebegun. Overall, households noted uniformly that re-cent climate change is already negatively affecting theirlives. The key factors identified by respondents that arecurrently affecting household livelihood activities andpose significant challenges to their future trajectory areshifting water availability, increasing weather extremes,and threats to tourism.

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

802 Mark et al.

Table 4. Climate change perception and new vectors ofhousehold vulnerability (95 percent confidence interval) inthe Querococha (QUER) and Quilcay (QUIL) case study

watersheds

Percentage

Perceptions and vectors QUER QUIL

Household perceptions of recent climate changeHouseholds reporting

significant recession ofnearby glaciers

100 94 ± 3.96

Households responding thatthey are very preoccupiedby recent climate changetaking place in the region

91 ± 4.3 84 ± 6.11

Households reporting thatthey have noted negativechanges in their lives due torecent climate change

94 ± 3.66 88 ± 5.42

New vectors of vulnerability affecting householdsHouseholds reporting that

water supplies during thedry season have beendecreasing

93 ± 3.61 81 ± 6.54

Households reportingsignificant changes inrecent weather patternsthat negatively affectedcrops or livestock

95 ± 3.61 100

Households reporting recentfreezing or extremeprecipitation events thatnegatively affected humanhealth, crops, or livestock

91 ± 3.61 100

Respondents reporting thatglacier recession already isor will negativelysignificantly affect tourism

100 (n = 27) 100 (n = 26)

Water constitutes one of the most important re-sources for households in the region for human con-sumption as well as agro-pastoral activities. Regularglacial discharge has historically provided perennialsupplies of water to households with access to landresources in the case study communities, but accord-ing to case study respondents this temporal regularity isshifting due to increasing local and regional demandson water supplies and because the magnitude of wa-ter resources during the dry season is declining. Acrossboth case study communities, more than 90 percent ofrespondents (QUER, 93 percent; QUIL, 94 percent) in-dicated that water supplies in the watersheds have beendecreasing during the dry season. Although less than5 percent of households in both communities argued

that they did not have enough water for human con-sumption, roughly one quarter of households (QUER,26 percent; QUIL, 22 percent) indicated that they didnot have enough water for irrigation. Households re-ported that these changes have had significant impactson agricultural and livestock productivity. Respondentsin both communities also noted that many of the peren-nial and intermittent springs in the watersheds havebegun to disappear. Households noted that diminishingspring flows in the upper reaches of the Querococha wa-tershed are negatively affecting livestock productivityin terms of health and growth rates, and that the de-cline or disappearance of springs in the lower reaches ofthe Quilcay watershed would have grave consequences,as 90 percent of case study households noted that theydepend on fewer than five springs across the entire areafor all of their potable water resources. Intervieweesalso uniformly indicated that one major spring abovethe community (located at ∼4,000 m) disappeared inthe past several years and that this has already affecteddry season water availability.

Another key vector of vulnerability affectinglivelihood activities in the case study communities isincreasing short-term weather variability. As Table 4illustrates, households across both communities reportthat intense precipitation events, freezing events,strong winds, shifting rainfall patterns, and intenseheat spells have all negatively affected householdhealth, agricultural productivity, and livestock health.Households also noted increasing interannual orseasonal variability in weather events and temperaturessuch as longer freezing periods, more damaging freezingevents, and shifts in crop planting or harvest periods.Respondents indicated that the cumulative effects ofthese increasing climatic extremes pose new risks forbasic household food security, are the source of greateruncertainty about agricultural cycles, and are oftenresponsible for substantial declines in crop productivity.

Finally, another important set of vectors affectinghousehold livelihoods are the effects of new climate-induced shocks on tourism activities in the region. Morethan one quarter of households in both case study com-munities engage in the provision of tourism services.Because the high glaciated peaks are one of the pri-mary reasons tourists visit the region, glacier recessionconstitutes a significant threat to tourism-related in-come for households. In fact, glacier recession has al-ready had significant consequences for households inthe vicinity of the Querococha watershed as the Pas-toruri glacier, which was visited by more than 40,000people in 2006, became the first glacier in the Cordillera

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

Climate Change and Tropical Andean Glacier Recession 803

Blanca to be closed due to “adverse climatic conditions”and is likely to completely disappear by 2015 (INEI2007).

Discussion

Glaciers are an integral component of the couplednatural–human systems of the tropical Central Andesand their rapid recession is transforming downstreamhydrology. Our integrated research has innovativelyquantified these changes and initiated new understand-ing of groundwater processes. Our isotope analyses showthat the relative amount of glacier melt water is increas-ing in glacierized tributaries of the Santa River, reflect-ing an increased rate of glacier recession in all the casestudy watersheds. Historically, a significant increase indischarge was observed until 1983, after which time thetrend is weakly negative. This inflection point in thedischarge trend is suggestive of a systemic threshold inglacier response to climate forcing, whereby glacierizedwatersheds initially provide more discharge but then di-minish in influence over time. In addition, temperatureis closely correlated to discharge, reflecting a regionalclimate control. It is important to note that 1983 wasa major El Nino year, and the role of El Nino south-ern oscillation has been identified as controlling dis-charge and glacier mass balance on multiyear intervals(Francou et al. 1995; Vuille, Kaser, and Juen 2008).Nevertheless, the overall decrease in discharge at LaBalsa is not altered by these episodic influences. Our re-sults demonstrate that groundwater is proportionatelyat least as important as glacier melt with respect to totalcurrent dry season streamflows, and as glaciers recede,the influence of groundwater, and its role as a seasonalbuffer, will become increasingly important.

Livelihood vulnerability in our case study watershedsis also being significantly affected by recent glacier re-cession. Our results clearly show that households areacutely aware of these changes and that new vectors ofvulnerability, including shifting water availability, in-creasing weather extremes, and threats to tourism, areaffecting household access to resources. Respondentsacross both watersheds uniformly indicated that thesefactors are significantly affecting their current liveli-hood activities. Because our methodology was designedto limit biases through the use of a random samplingframe, and our total household sample constitutes a sta-tistically significant representative population for bothwatersheds, our findings are highly suggestive that liveli-hood vulnerability has already increased significantly

and that there is a compelling need to address theseconcerns.

Our transdisciplinary findings suggest that there isan intriguing scale-dependent discontinuity betweenhousehold perceptions, on one hand, and what ourphysical measurements and models demonstrate on theother. Ongoing glacier melt in the Cordillera Blancatributaries is accompanied by increasing discharge andrelatively more glacier melt water contribution instreamflows from highly glacierized watersheds. How-ever, households from communities situated in bothof our glacierized case study watersheds indicate verystrongly that water supplies are declining, a trend that isonly clearly measured in the Santa River discharge fromthe entire Callejon de Huaylas. Although we are con-sidering a number of possible alternative explanations,given the minimal hydrologic influence of glaciers onthe entire Santa River basin (i.e., < 10 percent glaciercover), we hypothesize that increasing basin-wide waterwithdrawals from human activities might be an impor-tant explanatory factor.

With this report we present an initial synthesis ofresults from two of three case study watersheds andestablish the context for our ongoing research project.Exploring the way in which these observed patterns andhydrologic processes are currently interacting providesan important rationale for continued observations,model development, and testing and more intensivesocial research so that we might better integrate ourtransdisciplinary research with the project’s socialand scientific components and provide useful analysesto broader scientific and policy audiences. Finally,although this research focuses rather exclusively onhuman vulnerability, our larger research goals areintended to address household resilience and adaptivecapacity as well as larger governance questions affectingthe management of water resources throughout theSanta River watershed.

Acknowledgments

This research was funded by the National ScienceFoundation (NSF No. 0752175, BCS—Geography andRegional Science) and included a Research Experiencefor Undergraduates (REU) Supplement. Additionalfunding for the LIDAR flight came from the NationalAeronautics and Space Administration (NASA No.NNX06AF11G), The National Geographic SocietyCommittee for Research and Exploration, and theOhio State University Climate, Water & Carbon

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

804 Mark et al.

Program, and the Faculty Senate of the Universityof California, Santa Cruz. We acknowledge thecooperative assistance of Peruvian colleagues Ing.Ricardo J. Gomez, Ing. Marco Zapata, and others at theUnidad de Glaciologıa y Recursos Hıdricos in Huaraz,Peru, and the following interview research assistants:Carlos Torres Beraun, Oscar Lazo Ita, Erlinda MariluPacpac, Jesus Yovana Castillo, and Gladys Jimenez.We recognize REU undergraduates Laurel Hunt, SarahKnox, Galen Licht, Sara Reid, Michael Shoenfelt,Patrick Burns, Alyssa Singer, and Shawn Stone forfieldwork assistance. We also thank Kyung In Huh forassisting with LIDAR data display. This is Byrd PolarResearch Center contribution number 1396.

References

Ames, A., S. Dolores, A. Valverde, P. Evangelista, D. Javier,W. Gavnini, J. Zuniga, and V. Gomez. 1989. Glacierinventory of Peru, Part 1, 105. Huaraz, Peru: Hidrandina.

Baraer, M., J. M. McKenzie, B. G. Mark, J. Bury, and S.Knox. 2009. Characterizing contributions of glacier meltand groundwater during the dry season in the CordilleraBlanca, Peru. Advances in Geosciences 22:41–49.

Barry, R. G., and A. Seimon. 2000. Research for mountainarea development: Climatic fluctuations in the moun-tains of the Americas and their significance. AMBIO: AJournal of the Human Environment 29 (7): 364–70.

Bebbington, A. 2000. Reencountering development: Liveli-hood transitions and place transformations in the Andes.Annals of the Association of American Geographers 90 (3):495–520.

Bebbington, A. J., and J. T. Bury. 2009. Institutional chal-lenges for mining and sustainability in Peru. Proceedingsof the National Academy of Sciences 106 (41): 17296–301.

Bradley, R. S., F. T. Keimig, H. F. Diaz, and D. R. Hardy.2009. Recent changes in freezing level heights in theTropics with implications for the deglacierization of highmountain regions. Geophysical Research Letters 36.

Burton, I., G. White, and R. Kates. 1978. The environment ashazard. New York: Oxford University Press.

Bury, J. 2004. Livelihoods in transition: Transnational goldmining operations and local change in Cajamarca, Peru.The Geographical Journal 170 (1): 78–91.

———. 2005. Mining mountains: Neoliberalism, landtenure, livelihoods, and the new Peruvian mining in-dustry in Cajamarca. Environment and Planning A 37 (2):221–39.

Bury, J., B. Mark, J. M. McKenzie, A. French, and M. Baraer.Forthcoming. Glacier recession and human vulnerabilityin the Yanamarey watershed of the Cordillera Blanca,Peru. Climatic Change.

Byers, A. C. 2000. Contemporary landscape change in theHuascaran National Park and buffer zone, CordilleraBlanca, Peru. Mountain Research and Development 20 (1):52–63.

Carey, M. 2010. In the shadow of melting glaciers: Climate changeand Andean society. New York: Oxford University Press.

Cutter, S. 2003. The vulnerability of science and the scienceof vulnerability. Annals of the Association of AmericanGeographers 93:1–12.

Denevan, W. M. 1992. The pristine myth: The landscapeof the Americas in 1492. Annals of the Association ofAmerican Geographers 82 (3): 369–85.

Eakin, H. C. 2006. Weathering risk in rural Mexico: Climatic,institutional, and economic change. Tucson: University ofArizona Press.

Francou, B., P. Ribstein, H. Semiond, C. Portocarrero, andA. Rodriguez. 1995. Balances de glaciares y clima en Bo-livia y Peru: Impacto de los eventos ENSO [Climate andglacier balances in Bolivia and Peru: Impact of the ENSOevent]. Bulletin de l’Institut francais d’etudes Andines 24(3): 661–70.

Hewitt, K. 1983. Interpretations of calamity from the viewpointof human ecology. Winchester, MA: Allen & Unwin.

INEI. 2007. The 2007 national census: XI of population and VIof houses. Lima, Peru: Institute of National Statistics andInformation.

Intergovernmental Panel on Climate Change. 2007. Climatechange 2007: Impacts, adaptation and vulnerability. InWorking Group II Contribution to the IntergovernmentalPanel on Climate Change fourth assessment report, ed.M. L. Parry, O. F. Canziani, J. P. Palutikof, P. J. v. d.Linden, and C. E. Hanson. Cambridge, UK: CambridgeUniversity Press.

Juen, I., G. Kaser, and C. Georges. 2007. Modelling observedand future runoff from a glacierized tropical catchment(Cordillera Blanca, Peru). Global and Planetary Change59 (1–4): 37–48.

Kaser, G., I. Juen, C. Georges, J. Gomez, and W. Tamayo.2003. The impact of glaciers on the runoff and the re-construction of mass balance history from hydrologicaldata in the tropical Cordillera Blanca, Peru. Journal ofHydrology 282 (1–4): 130–44.

Knapp, G. W. 1991. Andean ecology: Adaptive dynamics inEcuador. Boulder, CO: Westview Press.

Leichenko, R. M., and K. L. O’Brien. 2008. Double exposure:Global environmental change in an era of globalization. NewYork: Oxford University Press.

Liverman, D. 1990. Drought impacts in Mexico: Climate,agriculture, technology, and land tenure in Sonora andPuebla. Annals of the Association of American Geographers80:49–72.

Mark, B. G. 2008. Tracing tropical Andean glaciers overspace and time: Some lessons and transdisciplinary im-plications. Global and Planetary Change 60 (1–2): 101–14.

Mark, B. G., and J. M. McKenzie. 2007. Tracing increas-ing tropical Andean glacier melt with stable isotopesin water. Environmental Science and Technology 41 (20):6955–60.

Mark, B. G., J. M. McKenzie, and J. Gomez. 2005. Hy-drochemical evaluation of changing glacier meltwatercontribution to stream discharge: Callejon de Huay-las, Peru. Hydrological Sciences Journal/Journal des SciencesHydrologiques 50 (6): 975–87.

Mark, B. G., and G. O. Seltzer. 2003. Tropical glacier meltwa-ter contribution to stream discharge: A case study in theCordillera Blanca, Peru. Journal of Glaciology 49 (165):271–81.

———. 2005. Evaluation of recent glacial recession inthe Cordillera Blanca, Peru (AD 1962–1999): Spatial

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011

Climate Change and Tropical Andean Glacier Recession 805

distribution of mass loss and climatic forcing. Quater-nary Science Reviews 24:2265–80.

Ministry of Energy and Mines. 2008. Electrical statistics 2008.Lima, Peru: Ministry of Energy and Mines.

Polsky, C. 2004. Putting space and time in Ricardian cli-mate change impact studies: Agriculture in the US GreatPlains, 1969–1992. Annals of the Association of AmericanGeographers 94 (3): 549–64.

Postigo, J. C., K. R. Young, and K. A. Crews. 2008.Change and continuity in a pastoralist community inthe high Peruvian Andes. Human Ecology 36 (4): 535–51.

Racoviteanu, A., Y. Arnaud, and M. Williams. 2008. Decadalchanges in glacier parameters in Cordillera Blanca, Peruderived from remote sensing. Journal of Glaciology 54(186): 499–510.

Sauer, C. O. 1941. Foreword to historical geography. An-nals of the Association of American Geographers 31 (1):1–24.

Turner, B., R. Kasperson, P. Matson, J. McCarthy, R. Corell,L. Christensen, N. Eckley, J. Kasperson, A. Luers, andM. Martello. 2003. A framework for vulnerability anal-ysis in sustainability science. Proceedings of the NationalAcademy of Sciences 100 (14): 8074–79.

Viviroli, D., H. H. Durr, B. Messerli, M. Meybeck, and R.Weingartner. 2007. Mountains of the world, water tow-ers for humanity: Typology, mapping, and global signifi-cance. Water Resources Research 43 (7): 13.

Vuille, M., B. Francou, P. Wagnon, I. Juen, G. Kaser,B. G. Mark, and R. S. Bradley. 2008. Climate changeand tropical Andean glaciers: Past, present and future.Earth Science Reviews 89 (3–4): 79–96.

Vuille, M., G. Kaser, and I. Juen. 2008. Glacier mass balancevariability in the Cordillera Blanca, Peru and its relation-ship with climate and the large-scale circulation. Globaland Planetary Change 62 (1–2): 14–28.

Watts, M. J., and H. G. Bohle. 1993. The space of vulnerabil-ity: The causal structure of hunger and famine. Progressin Human Geography 17 (1): 43.

White, G. F. 1942. Human adjustment to floods. Chicago:University of Chicago Press.

Young, K., and J. Lipton. 2006. Adaptive governance and cli-mate change in the tropical highlands of western SouthAmerica. Climatic Change 78 (1): 63–102.

Zimmerer, K. S. 1991. Wetland production and smallholderpersistence—Agricultural change in a highland Peru-vian Region. Annals of the Association of American Geog-raphers 81 (3): 443–63.

Correspondence: Department of Geography and Byrd Polar Research Center, The Ohio State University, 1036 Derby Hall, 154 North OvalMall, Columbus, OH 43210, e-mail: mark.9@osu.edu (Mark); Environmental Studies Department, University of California, Santa Cruz,Santa Cruz, CA 95064, e-mail: jbury@ucsc.edu (Bury); akfrench@ucsc.edu (French); Earth and Planetary Sciences, McGill University, 3450University Street, Montreal, QC, Canada H3A 2A7, e-mail: jeffrey.mckenzie@mcgill.ca (McKenzie); michel.baraer@mail.mcgill.ca (Baraer).

Dow

nloa

ded

by [

McG

ill U

nive

rsity

Lib

rary

] at

12:

26 2

8 N

ovem

ber

2011